Sociological research methods may be divided into two broad, though often supplementary, categories:[106]
Qualitative designs emphasize understanding of social phenomena through direct observation, communication with participants, or analysis of texts, and may stress contextual and subjective accuracy over generality.
Quantitative designs approach social phenomena through quantifiable evidence, and often rely on statistical analysis of many cases (or across intentionally designed treatments in an experiment) to establish valid and reliable general claims.
Sociologists are often divided into camps of support for particular research techniques. These disputes relate to the epistemological debates at the historical core of social theory. While very different in many aspects, both qualitative and quantitative approaches involve a systematic interaction between theory and data.[107] Quantitative methodologies hold the dominant position in sociology, especially in the United States.[35] In the discipline's two most cited journals, quantitative articles have historically outnumbered qualitative ones by a factor of two.[108] (Most articles published in the largest British journal, on the other hand, are qualitative.) Most textbooks on the methodology of social research are written from the quantitative perspective,[109] and the very term "methodology" is often used synonymously with "statistics". Practically all sociology PhD programmes in the United States require training in statistical methods. The work produced by quantitative researchers is also deemed more 'trustworthy' and 'unbiased' by the general public,[110] though this judgment continues to be challenged by antipositivists.[110]
The choice of method often depends largely on what the researcher intends to investigate. For example, a researcher concerned with drawing a statistical generalization across an entire population may administer a survey questionnaire to a representative sample population. By contrast, a researcher who seeks full contextual understanding of an individual's social actions may choose ethnographic participant observation or open-ended interviews. Studies will commonly combine, or 'triangulate', quantitative and qualitative methods as part of a 'multi-strategy' design. For instance, a quantitative study may be performed to obtain statistical patterns on a target sample, and then combined with a qualitative interview to determine the play of agency.[107]
Sampling[edit]
The bean machine, designed by early social research methodologist Sir Francis Galton to demonstrate the normal distribution, which is important to much quantitative hypothesis testing[a]
Quantitative methods are often used to ask questions about a population that is very large, making a census or a complete enumeration of all the members in that population infeasible. A 'sample' then forms a manageable subset of a population. In quantitative research, statistics are used to draw inferences from this sample regarding the population as a whole. The process of selecting a sample is referred to as 'sampling'. While it is usually best to sample randomly, concern with differences between specific subpopulations sometimes calls for stratified sampling. Conversely, the impossibility of random sampling sometimes necessitates nonprobability sampling, such as convenience sampling or snowball sampling.[107]
Methods[edit]
The following list of research methods is neither exclusive nor exhaustive:
Archival research (or the Historical method): Draws upon the secondary data located in historical archives and records, such as biographies, memoirs, journals, and so on.
Content analysis: The content of interviews and other texts is systematically analysed. Often data is 'coded' as a part of the 'grounded theory' approach using qualitative data analysis (QDA) software, such as Atlas.ti, MAXQDA, NVivo,[111] or QDA Miner.
Experimental research: The researcher isolates a single social process and reproduces it in a laboratory (for example, by creating a situation where unconscious sexist judgements are possible), seeking to determine whether or not certain social variables can cause, or depend upon, other variables (for instance, seeing if people's feelings about traditional gender roles can be manipulated by the activation of contrasting gender stereotypes).[112] Participants are randomly assigned to different groups that either serve as controls—acting as reference points because they are tested with regard to the dependent variable, albeit without having been exposed to any independent variables of interest—or receive one or more treatments. Randomization allows the researcher to be sure that any resulting differences between groups are the result of the treatment.
Longitudinal study: An extensive examination of a specific person or group over a long period of time.[citation needed]
Observation: Using data from the senses, the researcher records information about social phenomenon or behaviour. Observation techniques may or may not feature participation. In participant observation, the researcher goes into the field (e.g. a community or a place of work), and participates in the activities of the field for a prolonged period of time in order to acquire a deep understanding of it.[25]: 42 Data acquired through these techniques may be analysed either quantitatively or qualitatively. In the observation research, a sociologist might study global warming in some part of the world that is less populated.
Program Evaluation is a systematic method for collecting, analyzing, and using information to answer questions about projects, policies and programs,[113] particularly about their effectiveness and efficiency. In both the public and private sectors, stakeholders often want to know whether the programs they are funding, implementing, voting for, or objecting to are producing the intended effect. While program evaluation first focuses on this definition, important considerations often include how much the program costs per participant, how the program could be improved, whether the program is worthwhile, whether there are better alternatives, if there are unintended outcomes, and whether the program goals are appropriate and useful.[114]
Survey research: The researcher gathers data using interviews, questionnaires, or similar feedback from a set of people sampled from a particular population of interest. Survey items from an interview or questionnaire may be open-ended or closed-ended.[25]: 40 Data from surveys is usually analysed statistically on a computer.